When artificial intelligence takes a few days to create a work that takes a whole crew several weeks to complete, will the film and television drama industry, especially short dramas and advertising businesses, be frustrated? Recently, a short video of OpenAI's large-scale video

entertainment 1892℃

When artificial intelligence takes a few days to create a work that takes a whole crew several weeks to complete, will the film and television drama industry, especially short dramas and advertising businesses, be frustrated?

Recently, a short video of OpenAI's large video generation model Sora has sparked heated discussions. While people are seeing the development of technology, they can't help but worry about the future of actors, directors, Fu Huadao and other film and television industry practitioners. What level can the video generated by sora reach at the technical level? Can it replace jobs such as photographers, directors and actors?

When artificial intelligence takes a few days to create a work that takes a whole crew several weeks to complete, will the film and television drama industry, especially short dramas and advertising businesses, be frustrated? Recently, a short video of OpenAI's large-scale video  - Lujuba

"Delighted and panicked"

Recently, an online blogger released an animated short film "Journey to the West" generated using AI technology. It has a complete plot and rich details, which aroused heated discussion among netizens. The high-profile artificial intelligence video model Sora released by OpenAI can create realistic videos based on user text prompts. Guo Tao, a senior artificial intelligence expert at

, told China Business News that theoretically, if sora can effectively use multiple 1-minute videos to generate coherent, high-quality micro-short plays or movies, then it will be better in picture details, light and shadow, and color. The performance in other aspects may be very close to the works shot by professional photographers. This is because AI can learn and imitate a large number of professional photography works to achieve similar visual effects when synthesizing videos. However, actual performance also depends on the complexity of the sora algorithm and the diversity of training data. For the industry, this could mean cost reductions and improvements in production efficiency, while it could also spark new discussions about copyright and creative ownership. The emergence of

sora has indeed made some practitioners happy and panicked.

photographer Li Wenjia told China Business News: "I'm very worried about my job. Sora can make images that are hard to come by, have rare shooting opportunities, and cost a lot of money at zero cost. Even from a technical perspective, some of the parts produced by Sora In terms of picture details, light and shadow, color, light and shadow, props, character movements, etc., the video will be very close to the works shot by professional photographers."

Chinese young playwright and director Xiang Kai also said that sora can achieve some things that humans cannot shoot. difficult content screen.

film director Zhuo Lan believes that sora should be more helpful to friends who are up owners. When text is needed to generate a video, there is no need to find files from the material library. Especially the kind of mass-produced videos that require simple expressions. If used well, it can indeed greatly improve productivity and aesthetics.

There are also some more rational views in the industry. Director and producer Chen Liang said that sora has the technical possibility to produce short plays of 1 to 10 minutes, and even mid-length plays of 20 to 30 minutes. But if you want to make a 90-minute movie, there are extremely high requirements for the sora's computing power and the richness of the material library, and it is not so easy to overcome. This is currently only a theoretical possibility, so It is difficult for sora to complete this kind of leap in a short period of time.

When artificial intelligence takes a few days to create a work that takes a whole crew several weeks to complete, will the film and television drama industry, especially short dramas and advertising businesses, be frustrated? Recently, a short video of OpenAI's large-scale video  - Lujuba

sora is suitable for assisting creation, but lacks emotional expression

He Wenbin, who has been engaged in short video and short drama creation for a long time, told China Business News that judging from the several AI works currently presented, the production effect of sora is still good and is very suitable for assisting creation. , the rendering or shooting of some effects can be achieved through such AI technology, which can greatly improve efficiency and ensure quality with refined granularity. The key is to shorten the production time and save time and cost.

Liu Jun, vice president and content creative executive president of Unilumin Technology (300232.sz), one of the virtual shooting technology providers who interviewed and recorded the late-stage TV series "Flowers", told China Business News on February 20 that, just like a year ago The launch of chatgpt and sora is considered to be another milestone moment for AGI (general artificial intelligence). But this model is so sensational not just because the video generated by AI is longer and has higher definition, but because OpenAI has surpassed the capabilities of all past AIGCs and generated video content related to the real physical world.

"From text (chatgpt) to pictures (dall·e) to video (sora), agi continues to accelerate its coverage of the entire process of film and television production.Liu Jun said that UniMing Technology tries to analyze the pre-, mid-, and post-production stages of film and television production to find new opportunities.

In the early and creative stages of virtual shooting, gpt4 can assist in script generation, and sora can quickly generate storyboards. Video allows the storyline to be videoized in one step, which greatly saves the preparation period for script writing, original painting design, animation production, etc., shortening it from months to days. For directors, it is very intuitive to see the picture in the shortest time. In order to better adjust the script and picture effects, the script and picture special effects are pre-processed, and the time saved can be fully prepared for post-production work.

In the middle of the virtual shooting, the video generated by sora restored the real physical perspective, lighting, and water flow Physical effects such as these can be directly applied to the large background screen of virtual shooting, and combined with the camera to capture real-person movements and facial expressions, which improves the production efficiency of virtual shooting. "Sora has been able to produce high-quality videos, which can have a great impact on asset design and production. It has a great auxiliary effect, and designers can focus more on the design creativity itself, which makes it easier for studio creativity to become a bonus. "Liu Jun said.

In the late stage of virtual shooting and the post-editing stage of the studio, due to sufficient preparation and production in the early and mid-term, the post-editing workload can be reduced a lot. Combined with the ability to enhance AI content, it can even achieve fully automated editing and direct editing. "In general, the emergence of gpt4+sora has made virtual shooting easier and greatly lowered the threshold for film and television production. "Liu Jun believes that good film and television production is still inseparable from innovation and creativity, and good IP content will become precious.

Although sora technology has great potential, many people in the industry believe that it may still have some flaws. For example , AI may encounter difficulties in processing complex plots, delicate emotional expressions, or culturally specific elements. In addition, AI-generated videos may lack the unique creativity of human artists.

Zhuo Lan believes: "The AI ​​I understand is all from It is built on existing content, and the core requirement of film creation is innovation. "

Xiang Kai also has a similar view. In his opinion, the videos produced by sora still have problems such as frame skipping and flash points. The emotional expression of the characters is relatively blunt. The technology still does not fully meet the requirements of film and television levels. It will not be possible in the short term. Substitute the works shot by photographers.

"The videos produced by sora are not comparable to the film and television works shot by photographers. They are completely two concepts. No matter how sophisticated and perfect the videos produced by sora are, they are all based on material library analysis and calculations. Then, data analysis is performed after keyword extraction, and the existing material library is used for generation. "Chen Liang analyzed, "Because of this, I don't think there is an inevitable substitution relationship between the two. If there is a real threat, it may be a threat to the video display and advertising industries. "

Chen Liang said for example that the "Journey to the West" video generated by AI technology seems to be entertaining to a certain extent. "But when we look at it in depth, we will find it very lengthy because all the content is Analysis and calculations are performed based on existing materials. "

Group performances and basic editing may be replaced.

Xiang Kai believes that it is difficult to replace sora, such as screenwriter, director, photographer, lighting engineer and other jobs, because he does not have the ability to think, but he may put future production The cycle has been shortened from 10 years to 3 years. Wu Jia, a practitioner in the digital transformation service industry of

, believes that movies should still reflect the relationships and emotions between people. Today’s sora does not know these. "Whether it is "Hot and Hot" 》 or "The Wandering Earth", there are values ​​and consistency in the movie. Now you split the script into 1000 prompts (a form of input used to represent what actions the AI ​​model should take or what output it should generate when performing a specific task) , it is very difficult for you to rely on people to correct the 'inconsistencies' in the generated video. "

Wu Jia also said that virtual and reality will definitely be integrated in film and television, and there will be many creative ideas. For example, movies like "Avengers" that combine animation and the real world will be richer in the hands of directors.3D modelers and renderers may be unemployed, and extras may not be needed that much, but those veteran actors, directors, and screenwriters are unlikely to be replaced.

Guo Tao believes that some types of jobs with high repetitiveness and low technical content may be affected, such as some basic-level editing work. Future film and television production may be a model of cooperation between humans and AI, rather than a unilateral replacement.

In addition to directors and editors, real actors are the focus of everyone's attention.

Liu Jun told China Business News that sora can improve the efficiency of virtual shooting, but in the short term it will be difficult to replace the effects of real-person performances such as Mr. Bao in "Flowers".

In the short term, for Hollywood or teams with mature film and television production experience, sora can improve the efficiency of realizing specific scenes. "Live performances focus on the actors' delicate movements and expressions, similar to the effects of real actors such as Mr. Bao, Lingzi, and Uncle Ye in "Flowers". It is difficult to be replaced by Sora in a short time." Liu Jun said that machines can assist actors , allowing them to free up more valuable time from tedious work and do more interesting things. "Existing videos may be re-produced with the assistance of AI video, and virtual shooting can create better works with the empowerment of sora."

Liu Jun revealed that UniMing Technology has obtained Microsoft Independent Software Developer (ISV) certification, and Obtained the official access license of openai. Currently, Unilumin Technology can provide one-stop virtual shooting solutions through its virtual shooting system unistudio.

"The existence of real actors must be necessary. When we go to watch a film and television work, we also watch the actors' interpretations, not just the background pictures and special effects. There are also many viewers who go to watch films and television shows for their favorite actors. works. Then some people say that you can actually use AI to directly generate the image of your favorite actor to perform the work. Then there will be infringement and ethical issues involved. Not all content can be generated by AI. If Sora can invest heavily in the future Commercial use may affect the production of some short videos, short plays and advertisements, but it will not completely replace them, and the use of AI must pay attention to production risks and ethical issues." He Wenbin analyzed to a reporter from China Business News.

Regarding the future of sora, Chen Liang believes that it is a long process from the emergence of any technology to actual mass production, and no one can predict how high the cost will be.

As for whether sora may replace practitioners in the film and television industry, Chen Liang believes: "Whether it is producers, directors, actors, photographers and other types of work, the biggest difference between the film and television industry and other industries is that the work has the strong personal will and intention of the producer. Emotional tendencies are precisely the reflection of personal artistic level, and are also the core existence of a film and television work. I don’t think Sora will have such abilities. There are some things that technology can never achieve or replace."

(Wu Jia, in the text) He Wenbin is a pseudonym)

Tags: entertainment